DSO: A GPU Energy Efficiency Optimizer by Fusing Dynamic and Static Information

TL;DR

This paper introduces DSO, a GPU energy optimizer that combines static and dynamic information to improve energy efficiency by 19% with minimal performance loss, addressing limitations of existing DVFS-based solutions.

Contribution

The paper presents a novel theoretical energy efficiency model and a machine learning-based approach that jointly utilize static code features and runtime metrics for GPU power management.

Findings

DSO improves GPU energy efficiency by 19%.

Maintains performance within 5% loss.

Leverages both static and dynamic information effectively.

Abstract

Increased reliance on graphics processing units (GPUs) for high-intensity computing tasks raises challenges regarding energy consumption. To address this issue, dynamic voltage and frequency scaling (DVFS) has emerged as a promising technique for conserving energy while maintaining the quality of service (QoS) of GPU applications. However, existing solutions using DVFS are hindered by inefficiency or inaccuracy as they depend either on dynamic or static information respectively, which prevents them from being adopted to practical power management schemes. To this end, we propose a novel energy efficiency optimizer, called DSO, to explore a light weight solution that leverages both dynamic and static information to model and optimize the GPU energy efficiency. DSO firstly proposes a novel theoretical energy efficiency model which reflects the DVFS roofline phenomenon and considers the tradeoff between performance and energy. Then it applies machine learning techniques to predict the parameters of the above model with both GPU kernel runtime metrics and static code features. Experiments on modern DVFS-enabled GPUs indicate that DSO can enhance energy efficiency by 19% whilst maintaining performance within a 5% loss margin.